Globalizing plant knowledge beyond bioprospecting?

At 8 am on the first of September 2023, I find myself in Fumesua, a small town close to Kumasi, the second largest Ghanian city and the historical seat of the Ashanti kingdom. I am visiting the Crop Research Institute (CRI), the national center for plant and agricultural science since 1964, and I am taking a walk through their cassava field trials together with the agronomist and technicians in charge. The cassava root (Manihot esculenta, also known as manioc or yuca) is a key staple crop for Western Africa as well as Brazil and Indonesia. Cassava plants therefore take pride of place among the crops studied at CRI, with several fields hosting experiments that range from identifying sturdy, drought-resistant varieties to testing propagation and storage methods, verifying characteristics of varieties in demand for local markets, and finding ways to facilitate farmers’ everyday work.

As we stroll between lines of cassava trees, we spend some time discussing intercropping experiments: that is, the planting of two or more crops in the same site in an effort to ensure that farmers have more frequent harvests (cassava would otherwise only be harvested once per year), while also helping to enrich the soil with more nutrients and reducing the presence of weeds by keeping the space between cassava plants occupied with non-invasive plants. Legumes turn out to be a good companion for tubers such as cassava, since they do not seem to compete for nutrients, are very effective in suppressing weeds, and are beloved by farmers because of high demand in local markets. The agronomist points with great affection to a Canavalia species (a genus also known as jack bean), a gorgeous little flowering plant growing around the tall cassava stems.

Canavalia-cassava intercropping is the subject of a 5-year field trial, which is yielding promising results despite being only in its second year. Just opposite this experiment there is a field hosting 14 different cassava landraces acquired from local farming communities, which are being tested for their yield, vulnerability to disease, and the quality of their produce. The agronomist wishes she could harness this biodiversity within many field trials at the same time, for instance by testing the effectiveness of intercropping different types of legumes with different cassava varieties, but there are severe financial constraints on such efforts.

CRI has recently stopped getting governmental money for its research activities, needing instead to rely largely on external funding sources. These types of experiments are not necessarily popular with international funders—such as the Bill and Melinda Gates Foundation—which are more interested in supporting novel forms of precision agriculture aiming to enhance the yield and resilience of crops through data-intensive studies of their genetics and physiology. Such studies are typically carried out in laboratories equipped with cutting-edge genomic sequencing technologies and experimental fields utilizing semi-automated data collection tools such as robots and drones, with an emphasis on identifying crop varieties which can be developed on a large scale and marketed as appropriate to specific environmental conditions (Williamson and Leonelli 2022). By contrast, many of the methods being tested at CRI are resolutely low-tech, with little need for fancy tools to achieve promising results for local cultivation; and they are particularly well-suited to support sustenance agriculture, thereby prioritizing Ghanian communities over the demands of international markets.

These very circumstances make efforts to steward data from these experiments ever more important, both to maintain a memory of what has been learned and to circulate that information beyond CRI to peers nationally and internationally. There is a thirst for tools and funding that may enable better memory practices and communication flows between researchers and relevant peers, including farmers, policy-makers, agricultural experts, and so forth. Much of the knowledge being shared is local expertise emerging from farming communities, which largely consists of ethnobotanical knowledge concerning which species thrive in which environments, how to select for specific traits, how to successfully grow different variants, or what the uses and problems associated with each plant may be. Mediating between Indigenous naming systems and scientific classifications is also crucial to this work, since it enables researchers to bring ethnoscience to bear on Western approaches to crop production in ways that are expected to benefit local communities.

In this context, technology is expected to play a crucial role. The digitization of observations and measurements from phenotypic data such as those derived from CRI experiments is widely seen, both locally and internationally, as a way forward (Coppens et al. 2017); while the digitalization of the process of data collection is perceived as essential to interpreting the data and contextualizing them in relation to ethnoscientific knowledge (Bronson 2022).

“Digital is now the new trend,” I am told by the technician, who explains how he has just received training by the International Institute for Tropical Agriculture (IITA) in Nigeria to facilitate the digitalization of experimental and ethnobotanical data and their incorporation into open databases. A key innovation developed by this group over the last ten years has been online platforms to collect and upload images and observations directly from the field to the cloud via one’s mobile phone, such as the Agronomy Field Information Management System (AgroFIMS). This app aims to facilitate the flow of data from field to databases where experimental results can be compared across sites, while also saving researchers time from manual annotation and reporting of those records. After spending the last few years documenting the hard work underpinning the development of these tools in high-powered research centers with extensive access to cutting-edge laboratory and data science facilities (but relatively little direct interaction with prospective crop consumers),^[1]See Leonelli 2022a/b, and the account of the prospects for these tools by Coppens et al. 2017 and Arnaud et al. 2022. I am interested to hear that they are proving useful to researchers in locations such as CRI, where scientific experimentation on novel forms of cultivation is not necessarily supported by use of the latest research instruments and computational infrastructure, and researchers work in much closer proximity to local farming communities.

Such efforts to study crops in ways that create a bridge between professional researchers and other knowledgeable stakeholders in agriculture have a long history, much of which revolves around the quest for mechanisms for sharing insights and data across relevant communities (Curry 2022). A particularly prominent attempt consists in a 50-years-long collaboration between the Food and Agriculture Organization (FAO) and the Consultative Group for International Agricultural Research (CGIAR), two multilateral organizations that have long pushed for a global reform of the ways in which information about crops are collected and exchanged across territories and groups—including both scientific and ethnoscientific data (Curry and Leonelli 2023). This reform includes the development and implementation of tools like AgroFIMS, with the goal to bring together a comprehensive body of knowledge on crop varieties which could inform agricultural, medical, and environmental research around the world for decades to come. Such data are conceptualized as global commons of inestimable value in the fight against climate change—resources that could and should be accessible to anybody working towards a better understanding of crop biodiversity and its role in relation to agricultural development strategies, food production cycles, and all other human interactions with the plant world.

On top of the technologies and infrastructures, a huge institutional effort accompanies and supports the collection and dissemination of plant data as global goods, including open data policies requiring researchers to share information as soon as they obtain it, governance that facilitates data travel across contexts and borders, and financial incentives towards utilizing existing data. The view of plant data as global commons exemplifies a quest for cosmopolitan, multilateral, transdisciplinary communication, whose reach and significance transcends existing geographical barriers, digital divides, national borders, and structural inequities (Leonelli 2023). It is a subset of the datafication of human, social, and environmental interactions which has come to epitomize the 21^st century and the advances of artificial intelligence; its promise is to accelerate and smooth the path towards food security and planetary health, making it possible to address the concerns raised by climate change and the loss of biodiversity (Miles 2019).

Walking through the cassava fields tended by CRI, however, one wonders about the plausibility and practical implications of this vision. As CRI researchers confirm when I ask them to recount their experiences, collecting a wide variety of data about cassava—and doing it in a way that is mindful of farmers’ preferences and relies on their experiences—has undoubtedly gotten much easier in recent years. At the same time, the digital ecosystem within which cassava data are being globalized retains serious challenges.

For a start, how such digital resources fare as collaborative platforms beyond academia remains unclear. Even within the research world, relevant skills are distributed very unevenly and not all CRI staff members are aware of these tools and able to use them. For instance, the cassava agronomist who accompanies me, though she is senior to the technician, has not received the same training in digital data management; and as I realized through interactions with researchers based at similar facilities in Southern Europe and Western Africa, many feel uninformed and unaware of the opportunities emerging from applications of digital technologies. While CRI management is committed to improving computing and data management skills among their staff, the institute does not have the resources to provide in-depth data training for all its staff or guarantee long-term data stewardship beyond the span of funded projects.^[2]The tension between data management aspirations and practical challenges is evident in a survey that Joyce Koranteng-Acquah and myself, in collaboration with CRI management, conducted on CRI researchers. Results highlighted scientists’ wish to improve their data skills, while also noting their difficulties in taking advantage of national and international digital infrastructure developed to support this type of work (Koranteng et al. 2024). Even when such resources are available, they need to be highly tailored to the specific crops and goals with which different research teams work. Unsurprisingly, what may work for cassava roots will not work for banana trees, and tools used to seek high-yield crop varieties may not be adequate to assess which varieties are most likely to be favored by local farmers and markets. Thus, the data-sharing landscape may vary considerably depending on the investment directed towards specific crops and socio-economic contexts, with findings coming from a few well-resourced loci used to generate knowledge that is then presumed to apply much more widely.

Moreover, the growth of crop data infrastructures has not resulted in increased recognition of the role played by local communities in creating such a knowledge base. Once data are extracted from local circumstances and placed online, they can be mined by any interested party—including large corporate agrobusinesses and powerful North American and European universities—with little acknowledgment and no credit to those who generated the data in the first place or to the farming communities who may have contributed expertise and prime materials (such as plant cuttings of local varieties) to the experiments (Marks et al. 2023).

Ethnoscience is thus recognized to the extent that it is extensively mined and appropriated, but—in unsurprising continuity with the colonial history of the Green Revolution—without due credit to its originators. Debates around who should reap the benefits of data-intensive crop research—and particularly whether traditional communities should partake in its development and profits—continue with little progress (Bronson 2022; Leonelli under review). While the history of efforts to secure fairness in germplasm exchanges goes back to the turn of the century, culminating in the successful ratification of the International Treaty on Plant Genetic Resources for Food and Agriculture in 2001, equivalent efforts to ensure fair trading of digital information—exemplified by the Global Information System supposedly associated with the implementation of the Treaty—lag behind.

The introduction of data technologies has also not challenged fundamental assumptions around who controls, governs, and profits from data-intensive crop research and precision agriculture. In fact, it may have further entrenched the uneven power dynamics and structural injustice already permeating the agricultural sector. Despite the insistence on collecting various forms of data, including morphological and socio-economic information, molecular data are consistently treated as more reliable than observational accounts of the physiology and development of morphological traits—let alone their social value to local communities—and therefore used as reference point to organize databases and data analysis.

In short, crop science still trumps ethnoscience. This has dramatic effects on whose expertise is valued and whose is not. Contributors with no genomic expertise and related instruments tend to be viewed as second-class contributors to agricultural research, and to not be formally credited or rewarded. Even the emerging legal frameworks around agricultural data are chiefly preoccupied with digital sequence information (in other words, molecular data about plant genomes) and its use towards identifying elite breeds. This in turn facilitates a “gene fetishism” that is dangerously aligned with centuries-old colonial understandings of biodiversity as a resource to be extracted and exploited rather than managed sustainably (Bonneuil 2019; Fenzi and Bonneuil 2016). It is also likely to result in further marginalization of farmer knowledge and local agrodiversity (Tarjem 2022), thereby diminishing the status and legitimacy of ethnoscience precisely at a time where its value is recognized in practice.

When considered within the broader landscape of commercialized datafication—where large corporate firms dominate the global market for data acquisition and data subjects have little capacity to participate in the development and governance of data infrastructures (Jensen 2020)—it becomes clear that the vision of crop data as global commons to be freely shared does not necessarily support collaborative research towards sustainable food production. In many instances it is being used as a sophisticated form of bioprospecting (Hayden 2005) and surveillance (Miles 2019), whereby local knowledge of crops and their uses is harnessed and mined in ways that damage the very communities that produced it.

In the case of cassava, a crop uniquely suited to sustenance agriculture (Rival and McKey 2008) and “life in the margins” (Bray 2021), this process is just starting. It is evident in the ongoing tensions between the attempt to support community farming of local varieties and the quest for mass scale production and processing of high-yield varieties, two rather different goals which uncomfortably co-exist within the data infrastructures set up to collect and disseminate cassava data across experimental sites. And it is reinforced by the lack of resources available to researchers wishing to explore sustainable uses of agroecology, such as the experimental testing of intercropping at CRI and their methods of engagement with local farmer communities. Efforts to foster understandings of local agrodiversity, and agricultural interventions grounded on such understandings as well as ongoing engagement with local communities and their needs, are among many promising avenues out of a purely exploitative datafication of ethnoscience (Vijayan et al 2022), and providing visibility and resources to those involved would go a long way towards facilitating a fairer and more sustainable approach to plant science and its societal uses.

Acknowledgments

This research has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No. 101001145, project “A Philosophy of Open Science for Diverse Research Environments”; www.opensciencestudies.eu ) and the Alan Turing Institute under the EPSRC grant EP/N510129/1. This paper reflects only the author’s view and that the Commission / Agencyis not responsible for any use that may be made of the information it contains. I am grateful to Joyce Koranteng-Acquah for prompting me to collaborate with CRI and her subsequent ethnographic work in the institute, which will soon appear in print as part of her PhD dissertation; to CRI management and staff for their hospitality and willingness to engage with our research; and to Hugh Williamson, Helen Curry and Rachel Ankeny for relevant discussions.

Works Cited

Arnaud, Elisabeth et al. 2020 “The Ontologies Community of Practice: An Initiative by the CGIAR Platform for Big Data in Agriculture.” Patterns 1: 100105.

Aubry, Sylvain. 2023 “Genbanking Plant Genetic Resources in the PostGenomic Era.” Agriculture and Human Values 40: 961–971. https://doi.org/10.1007/s10460-023-10417-7

Bonneuil, Christophe. 2019 “Seeing Nature as a ‘Universal Store of Genes’: How Biological Diversity Became ‘Genetic Resources’, 1890–1940.” Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences 75:1-14. https://doi.org/10.1016/j.shpsc.2018.12.002

Bray, Francesca. 2021 “‘Underground Inspirations: Tuber Sciences and Their Histories.’” Isis 112(3): 548-563. https://doi.org/10.1086/715436

Bronson, K. 2022. Immaculate Conception of Data: Agribusiness, Activists and Their Shared Politics of the Future. McGill University Press.

Coppens, Frederick, Nathalie Wuyts, Dirk Inzé, and Stijn Dhondt. 2017. “Unlocking the Potential of Plant Phenotyping Data through Integration and Data-Driven Approaches.” Current Opinion in Systems Biology 4: 58-63. https://doi.org/10.1016/j.coisb.2017.07.002

Curry, Helen Anne. 2022. Endangered Maize: Industrial Agriculture and the Crisis of Extinction. University of California Press.

Curry, Helen Anne and Sabina Leonelli. 2024. “Crop Descriptors and the Forging of ‘System-wide’ Research in CGIAR.” In Research as Development: Historical Perspectives on Agricultural Science and International Aid in the CGIAR Era, edited by Helen Anne Curry and Timothy W. Lorek. Cambridge University Press.

Fenzi, Marianna and Christophe Bonneuil. 2016. “From ‘Genetic Resources’ to ‘Ecosystems Services’: A Century of Science and Global Policies for Crop Diversity Conservation.” Culture, Agriculture, Food and Environment 38: 72–83. https://doi.org/10.1111/cuag.12072

Hayden, Cori. 2005. When Nature Goes Public: The Making and Unmaking of Bioprospecting in Mexico. Princeton: Princeton University Press.

Koranteng-Acquah J., Acheampong, P., Asante M., Baafi E., Mociah M. B. & Leonelli, S.* (2024) Data Management at the Crops Research Institute of the Council for Scientific and Industrial Research, Ghana: A Scoping Report. PHIL_OS Report, June 2024. DOI: https://doi.org/10.5281/zenodo.11479135   

Jensen, Jakob L. 2020 The Medieval Internet: Power, Politics and Participation in the Digital Age. Emerald Publishing Limited.

Leonelli, Sabina. 2022a. “How Data Cross Borders: Globalizing Plant Knowledge through Transnational Data Management and Its Epistemic Economy.” In Knowledge Flows in a Global Age: A Transnational Approach, edited by John Krige, 305-32. Chicago: University of Chicago Press.

Leonelli, Sabina. 2022b. “Process-Sensitive Naming: Trait Descriptors and the Shifting Semantics of Plant (Data) Science.” Philosophy, Theory and Practice in Biology 14:16. https://doi.org/10.3998/ptpbio.3364   

Leonelli, Sabina. Under review. When More Knowledge Is Less Power: Extractive Epistemologies in Crop Data Science. ST&S.

Marks, Rose A. et al. 2023. “A Critical Analysis of Plant Science Literature Reveals Ongoing Inequities.” PNAS 120(10). https://www.pnas.org/doi/10.1073/pnas.2217564120

Miles, Christopher. 2019. “The Combine Will Tell the Truth: On Precision Agriculture and Algorithmic Rationality.” Big Data & Society 6(1). https://doi.org/10.1177/2053951719849444  

Rival, Laura and Doyle McKey. 2008. “Domestication and Diversity in Manioc (Manihot Esculenta Crantz Ssp. Esculenta , Euphorbiaceae).” Current Anthropology 49(6): 1119-1128.

Tarjem, Ida Arff. 2022. “Tools in the Making: The Co-Construction of Gender, Crops, and Crop Breeding in African Agriculture.” Gender, Technology and Development 27(1):1-21. https://doi.org/10.1080/09718524.2022.2097621

Vijayan, D., Ludwig, D., Rybak, C., Kaechele, H., Hoffmann, H., Schönfeldt, H. C., … & Löhr, K. 2022. Indigenous knowledge in food system transformations. Communications Earth & Environment, 3(1), 213.

Williamson, H. and Leonelli, S. 2022. Accelerating Agriculture: Data-Intensive Plant Breeding and the Use of Genetic Gain as An Indicator for Agricultural Research and Development. Studies in the History and Philosophy of Science 95: 167-176. https://doi.org/10.1016/j.shpsa.2022.08.006

Williamson, Hugh and Sabina Leonelli, eds. 2023. Towards Responsible Plant Data Linkage: Global Data Challenges for Agricultural Research and Development. Cham: Springer Open Access.