Glimpsing the Cosmos in the Gaps of Optimisation

Human comprehension of the physical processes and properties of the Earth relies heavily on technology-driven systems like satellite imagery and climate models. These technologies have become essential for monitoring the impact of human actions on the planet. However, there exists a disconnect between local experiences and the abstract, algorithm-based climate projections that are beyond human perception. The need to bridge this gap is probed by Guillemette Legrand’s project ‘From Here to the Cosmos: Incomputable Views of the Above, Under and Around’. 

Legrand is an artist and designer who uses machine-fictioning to explore the (im)materiality of emerging information technologies and their capacity to form contexts that normalise specific belief systems. For this project, they worked together with Marian Neagul, who is currently a postdoctoral researcher at the Institute eAustria Timișoara and West University Timișoara. Neagul’s general research topics cover machine learning, distributed systems, computer networks, operating systems, and Earth observation. 

In this interview, Legrand discusses the concept and evolution of the project, what makes for valuable art-science collaborations, and the importance of recognising the ideologies and gaps in technological systems.

‘From Here to the Cosmos: Incomputable Views of the Above, Under and Around’ was one of the projects presented in the Turn Signals—Design is not a Dashboard exhibition. How was the project experienced? 

Guillemette Legrand: Visitors to the exhibition experienced the project as a fragment of what may be the inside of a sphere mesh; an archetype of the Earth’s shape. Standing in front of this fragment, visitors observed various algorithmic simulations of the geophysical condition of the Earth, produced by autonomous technical systems. These simulations were rendered at different timescales extracted from three environments of Timișoara: above (cloud detection), around (land usage detection), and under (ground motion monitoring). A voice-over renarrates the synthetic visuality and the computational logic behind what is seen, while longing for new forms of embodied and situated imaginations of the planet.

The project does indeed have an expansive ffect from the scale of personal observation to planetary complexity. What was the motivation behind the project? 

GL: I began with the intention of merging my current research on the practices and discourses around Earth and climate imaging, with the socio-cultural and scientific context of the city of Timișoara. To bring these two aspects together, I was interested in understanding the specific practice of ‘Earth-imaging’ in the context of an Earth-observation laboratory that is working with visualisation practices in Timișoara and Romania as a whole. The project is an attempt to re-narrate the technical systems that mediate the relationship between human, machine and Earth, by collapsing different timescale and synthetic visions, while searching for other ways to enrich and diversify the machine gaze from a situated and embodied perspective. The final installation is an attempt to create a situation where the viewer faces the land above, around and under their feet through the mediation of the technological gaze; in order to both understand and see beyond this gaze.

Why is seeing beyond the technological gaze important?

GL: The human understanding of the geophysical Earth is increasingly mediated by a complex infrastructure that senses and models the planet as a volumetric object, and that produces important knowledge determining the future of the planet. This technological mediation is the dominant gaze through which humans have been able to render visible the impact of human activity on the Earth and its climate. The visual culture emerging from this infrastructure operates at a more-than-human timescale; beyond human senses. This technical infrastructure has become the necessary proxy to maintain life on Earth, but this visuality has also created a gap between empiric and local experience, and the algorithmic vision anticipating climate collapse. The project emphasises that this visuality is insufficient knowledge, yet at the same time, this visuality also naturalises specific socio-cultural beliefs about our human relationship with the Earth. Consequently, our cosmological understanding of the world-to-come is shifting.

How did you come to collaborate with Marian Neagul and the other scientific partners involved in this project? 

GL: I first learned about Neagul’s research online, and contacted him to establish an initial conversation about his current work in ML-aided Earth observation and visualisation practice around land usage in Timișoara and Romania. This initial conversation with Neagul led to a series of informal interviews with Romanian scientists working on tools used to sense and model different Earth environments. These conversations were instrumental in developing an understanding of specific technical developments in Romania relating to planetary-scale infrastructure, such as the Copernicus programme—the Earth observation component of the European Union’s Space Programme. Apart from Neagul, there were two other key partners that I collaborated with: Terrasigna, who specialise in ground motion monitoring; and the Meteorological Office of Bucharest. Together, these three partners provided data and support in developing compelling visualisations to communicate Timișoara’s specific geophysical conditions across various timescales. The choice of these partners was responding to the idea of building a planetary collage of different machine gazes—from the skies, the Earth’s crust and the underground. The cooperation was developed with the idea of creating interfaces for these three layers of vision; from the cosmos to the underground, portraying different views of Timișoara.

How did you integrate this research into the project’s design process?

GL: In my work, I examine if and how machines and their operative systems fictionalise reality and naturalise specific perspectives that have the potential to push specific world-narratives or cosmologies. By appropriating the potential of algorithmic systems to fictionalise reality, I aim to create a different embodiment of this gaze, enabling different world-imaginations. 

In parallel with the conversations with scientists, I used such a machine-fictioning practice to understand both the socio-cultural imagination created by the existing technical apparatus, and re-narrate other stories of planetary visions experienced through a situated and embodied knowledge. At an early stage of the project, the iteration of the designed environment for the exhibition led some of the research direction. The shape of the projection surface references the grid systems used to divide, map and model the Earth. The lines of the grid have become voids; spatial interstices where knowledge can be imagined from different viewpoints, scales and beliefs.

How did this collaborative process make you reflect on your own domain?

GL: In my opinion, for an art-science collaboration, or indeed any collaboration to be meaningful, there always needs to be some time for genuine interest on both sides to develop. In art-science collaborations, it is often the case that an artist or designer’s discipline is seen as being to visualise or even aestheticise the work of the scientist. And it is true that in many cases, that is the easiest starting point for such a collaboration. 

At the same time, I am looking for other ways to be involved in the work of science. In this case, that is through trying to engage with earlier stages in the simulation of Earth by questioning the parameters and goals, the research output, and the belief involved in the technical processes behind these simulations. 

I also think that it is often expected for art-science collaborations to be productive in the sense that they need to have a material and clear outcome. But a lot of tacit, unspoken exchanges happen through engaging with practices from other disciplines. This can shape how one thinks about and practises one’s work, which to me is just as valuable. 

Can you speak more to the role of the visual, in both design and science, as a means of negotiating knowledge gaps? 

GL: I think this collaboration was essentially two visual practitioners reflecting on their way of reading, analysing, and producing content with very different languages to discuss and materialise these images. There were often misunderstandings surrounding what we thought was a ‘good image’. For example, from Neagul’s perspective an image without any clouds in it is an optimal image, because these images have more value by giving more information about different environments, such as land usage, or forestry. I, however, was interested in the clouds, and wanted to work with the rawer images—those without any additional alteration. This was because the rate of change between these images was greater, so the effect of interpolation in the video would be more readable and conceptually closer to what the project was trying to say. 

A lot of things that were discussed orally were misinterpreted on both ends due to linguistic and disciplinary differences. It was only when we started working through the actual process of exchanging images and visual content that the work really unfolded, and it became clearer what the project would look like.

The aesthetics of simulation suggest a logic of optimisation. How do you view the theme of optimisation—or lack thereof—in your own work?

GL: Optimisation is an inherent part of both human and machine decision-making, and it can be implicitly inherent to the practice of simulation. Through optimising, information is equally enhanced and suppressed, and that was one of the theoretical and technical starting points of the project. By understanding what is optimal for the system being dealt with, the beliefs and ideology woven into the system itself can be understood. 

As mentioned, in the installation there is a voice-over. This voice-over asks what is optimised or incomputable and therefore what information is lost and what information is created. This is a questioning of the bird’s-eye view synthetic visuality of the Earth. The voice-over goes on to ask: what other ‘optimised’ visuality could help reframe the current mediation of the technological gaze on our understanding of the Earth?