This is the story of a research project that is so wild that no funding agency would have approved it—but we believe in it, so we are doing it anyway. My colleague Meren said we should record the story of this project as oral histories. He already started his, this is the start of mine. For about three years I will try to regularly update this page with the latest developments from the project.
HUMAN PROGRESS
Today’s science is deeply fractionated, with continued progress requiring ever more specialization. At the same time there is a pressing need for bridging the gaps between fields and disciplines, to find cohesive solutions to the twin crises of climate change and biodiversity loss. However, the deep divide between the Sciences and the Humanities proves hard to overcome, although environmental research is hitting more and more boundaries where a deeper understanding of humans and human behaviour is needed: The research itself is done by humans with human biases, it describes environments that are impacted by humans and human decisions, and the products of this research are papers present the results as narratives for human audiences. The HIPP cohort HUMAN PROGRESS (in the sciences) aims to address the need for a deeper understanding of humans in the natural sciences, and specifically in marine environmental science and biodiversity research. The aim of this project is to learn from the humanities with a clear aim to solve natural science challenges. The project will intensify and upscale ongoing interactions between researchers at HIFMB and partners at the University of Oldenburg, to build lasting interdisciplinary teams that form a solid foundation for future progress.
Interdisciplinary Challenge
Successful interdisciplinary projects require more than finding a common language. The harder task is agreeing on what research is worthwhile, what constitutes a solution, and what is admissible evidence. Every discipline cherishes its past successes and thus successful methods and approaches are continually reused. Therefore researchers tend to focus on questions where established tools of their discipline can be applied. In the long run this makes it easy to deceive oneself into thinking that all important challenges have this form. When working with other disciplines it is then easy to disregard the alternative approaches they offer, because they don’t address the challenges that we are used to considering, and pay attention to concerns that we have long chosen to ignore. We believe that such divides can be overcome if we learn to understand the metrics of success of other disciplines, address issues openly, and unite to achieve important concrete goals that all involved disciplines recognize as worthwhile. The protection of marine biodiversity offers an abundance of such challenges.
Rules for Success
The PIs that will advise the HUMAN PROGRESS cohort have agreed on a set of common overarching goals:
- Our goal is to learn from the Humanities. We do not seek to outsource tasks to collaborators, but rather to enrich the natural sciences by incorporating methods and approaches that the humanities have established.
- Our metric of success is impact in the Sciences. While advances in the Humanities will be an added benefit. The central goal of the project can only be achieved by publications in high-impact journals in the Sciences.
- Results must speak to the Sciences. We embrace methods and approaches from a wide range of disciplines, but results must meet evidence standards from the Sciences, which may require stronger quantification and scaling of methods from the Humanities using modern computational approaches.
By embracing these rules as a team we are laying an important foundation for a successful collaboration that will benefit all partners.
Projects
Natural Scientists at HIFMB have identified five areas where input from the Humanities is needed. Each of these areas will be studied by a postdoctoral researcher from the humanities with the help of colleagues from the University of Oldenburg and in one case Tallinn University.
1. Names have power
The Ecologically and Biologically Significant marine Areas (EBSAs) are prospective MPAs that could contribute to the 30-by-30 goal of marine protection. But among these prospective areas only 28 reference any biology or biological entities in their name. By comparison 97 EBSAs reference sea floor features. These names hint at the thought processes that have led to the presently proposed areas and thus merit analysis. This analysis could profit strongly from Linguistics where Hydronomy, the naming of bodies of water, is an active field of research. But the EBSAs are only one example in which naming intersects with biological research. Because of the enormous complexity and diversity of biological systems, Biology is constantly in need of new names. These names reveal perceptions at the time of naming, but also consolidate nascent fields, shape expectations, and set agendas.
2. Historical Observations
In recent years environmental research has profited from the analysis of big data sources, which focused on structured data, such as tables of curated observations. However, there is still a considerable knowledge gap regarding the response of biodiversity to climate change as data tends to be recent (e.g. spanning decades) or from the fossil record (spanning millions of years with gaps lasting hundred thousands of years.) and thus data covering the timescale of climate change is still sparse. Meanwhile a wealth of less-structured data spanning centuries has been left untapped. This data comes in the forms of historical accounts and indigenous knowledge. Hence a logical next challenge is to enable large-scale computer-aided analysis of such data. This work can profit from recent advances in hand-writing and speech recognition and large language models, but further steps in the knowledge creation pipeline consisting of computer-aided systematization and canonization of data is still missing. For data, a natural middle ground between structured and unstructured representation is offered by knowledge graphs, i.e. networks in which data entities, such as people, places or objects are represented by network nodes, while network links encode their relationships. Knowledge graphs haven’t yet received much attention in network science, due to their semi-structured nature. Therefore the development of new analysis methods is necessary, which can draw inspiration from established non-computational workflows in historical research.
3. (Re-)examining Modelling
One of the pressing societal issues we face today is the loss of trust in science, which is particularly evident in the area of modelling. To address this loss of trust we must understand its origins. Presently mathematical models are mostly black boxes that are only understood by a few experts. Others are not in a position to meaningfully question these models and are thus left in a position where the only options are to accept the models and their implications or to reject them entirely. Modellers are aware that modelling necessarily involves simplifications that are not without alternatives, such that the choices made reflect the modelers ideas and biases and hence should be questioned. Indeed, a major way in which models generate knowledge is by enabling a critical questioning of assumptions. However, this power of advancing through questioning is presently only accessible to a small group of modellers. Efforts to reduce the distrust of models must therefore aim to create a meaningful dialogue that enables a wider base of people to question models, without needing to understand their mathematical basis. We believe that such a dialogue can be enabled by gaining a philosophical understanding of how models produce insights and by communicating these pathways of insight creation directly and openly.
4. Human decisions
In mathematical models we typically model ecological systems as if humans did not exist. For example a large majority of published ecological food webs do not include humans. By contrast there is hardly any system left on the planet that is not severely impacted by humans. When we incorporate humans into models, e.g. in the context of socio-ecological system models, we typically treat the humans as an additional predator—the fisherman becomes indistinguishable from a shark. Modelling humans in this way ignores a wealth of knowledge on human behavior and decision making, including psychological, social, political, and economic dimensions. Taking these additional complexities into account and properly incorporating them in models has the potential to shift our perspective on the ecological dynamics.
5. Science narratives
Even in abstract mathematics, the products of research are papers that present human-readable narratives. In the environmental sciences new narratives appear to be relatively rare, and seem to emerge mostly from agenda-setting papers in leading journals, while most publications restate and reinforce existing narratives. The emergence of new narratives are thus major events that can shift the focus of research. However, while narratives are commonly studied in the context of novels, short stories and plays, there is almost no work on the narratology of science. We can ask for example, what is the structure of successful scientific narratives? What are the circumstances in which new narratives can be established? Are there research themes that are not pursued because they don’t make good stories? What is the relationship between the emergence of narratives and the emergence of new research fields? And, how do scientific narratives change when they are communicated to the general public?
Overarching concept
Collectively, the five projects trace the stages of discovery, from establishing a basic ontotology, via data analysis and modelling to narrativization and decision making. On a different level, the project also addresses obstacles faced by interdisciplinary research itself, including the emergence of diverging terminology, narratives, models, and methods. On a third and perhaps most important level the projects address five areas where a better understanding of humans could aid biodiversity research and biodiversity conservation: overcoming data limitations by tapping less-structured data sources, understanding frictions in interdisciplinary research, improving the representation of humans in environmental models, and overcoming obstacles in the communication of results within science and to the general public.