One of the driving forces behind this exhibition – where photography, neuroscience, mathematics and artificial intelligence meet – is a hybrid approach, used to give cause for experimentation and to remind us, in passing, of photography’s privileged status in the history of science. Taking the notion of latent space as their starting point, the photographer Marina Gadonneix (a 2002 graduate of the ENSP photography school and the 2006 winner of the HSBC Prize for Photography and 2020 winner of the Niépce Prize) and the artificial intelligence scientist Victor Rambaud (a cognitive science researcher and École Polytechnique graduate) explore possible representations of the mind and open up stimulating reflections on the status of images.
Latent space, a central notion in AI, denotes the hidden structure through which broad sets of data are organised. In this notion, Marina Gadonneix found the ideal starting point for questions that cast light on the potential of images, on their grey areas and on the polarity that runs through them in a continuation of her reflections on the production and hidden dimension of images and on the simulation and appearance of phenomena. These continued reflections led her to the heart of science laboratories (Phénomènes, 2019; Laboratoires/observatoires, 2023).
It was, naturally, on this fertile ground that she met Victor Rambaud – a legacy of those dialogues between artists, scientists and engineers that took place in the late 1960s, such as Experiments in Art and Technology and the Center for Advanced Visual Studies at the Massachusetts Institute of Technology. For the scientist, strongly aware that AI is, in the collective imagination, often confined to science fiction and dystopia, it was about exploring an unexpected aspect of AI in its relation to cognitive science and the mind. Through the multiplication of approaches, it was also about producing a more comprehensive perception, while bringing out its poetic dimension too.
This shared desire for experimentation was proportionate to the breadth and complexity of the subject. Even though several guiding lights – including da Vinci, Bachelard, Musil, Thom and Morin – are touched upon, such polymaths and interdisciplinary luminaries are only evoked in order to better invite us to grasp the conceptual, mouldable promises of those abstract spaces and, more generally, the way in which scientific representations help us understand the way in which the mind works.
An iconographic atlas – a tribute to Aby Warburg’s Mnemosyne Atlas – introduces the exhibition’s theme and reveals its full wealth. From one room to another, the images are in harmony with each other. From antique myths to neural networks, echoes shed light on attempts to represent the most abstract forms of thought. Topology – “surveying through the mist” in the words of the genius mathematician Alexander Grothendieck – here takes the shape of daydreaming that leads us to neuronal valleys. In the work Constellation, words themselves become images through projection in latent spaces, opening up a poetic, expressive surface upon which each person can display their own visions of the mind. The scientific pathway designed by Victor Rambaud accompanies this exploration implicitly and underlines that the history of mental representations has always inextricably mixed perceptibility with abstraction.
The point of convergence between cognitive science and artificial neuronal networks is based on recent discoveries in line with work by O’Keefe and Moser (2014 Nobel Prize) on spatial representations in the brain. It casts light on a common geometrical foundation through topological varieties such as the torus – intriguing surface forms where the inside and the outside merge. To understand them, Marina Gadonneix turned to the mathematical models at the Institut Henri Poincaré that she photographed in a series that presents them in plain sight, straying away, in terms of form, from the paths trodden by Man Ray (Mathematical Objects, 1934–1936) and Hiroshi Sugimoto (Mathematical Forms, 2004).
In the manner of those tangible abstractions, the photographer took an interest in other objects with uncertain statuses that evoke representations of the mind, whether sculptures of latent spaces (the series Walking down my soul) or discreet shadows of these objects (Contre-forme). Enigmatic and suggestive, these photographs reveal uneven reliefs where mountains take shape in hollows or rises, helping us view an in-between space where inner and outer landscapes merge. The aesthetic stances here create a spectacular play between the different series that eludes the oppositions of real versus virtual, material versus immaterial, true versus fake, meaning versus presence and so forth, strengthening the unity of the whole.
Over the course of exhibits, the different series by Marina Gadonneix respond to each other and extend a framework that goes beyond documentary intention. Géométries de l’esprit (Geometries of the mind) pursues a path towards abstraction and fundamentals, where the photographer’s recurring questions are amplified and matrixed by the attention given to the black box of the mind. After Landscape/Blackout (2011), this is about taking away the decor and bringing forth possibilities. After the series Après l’image (2014–2016) and Phénomènes (2019), it is about evoking simulation of reality and construction of images – yet while turning here to a complex reality, examined down to its most intimate lines, where the system (that exterior of the image, vital to showing it) gives way to the structure (that interior of the image, the condition for its wholeness).
Through this investigative work on the matrix of our mental representations, the exhibition achieves the feat of producing a mise en abyme of questions about photographic images, where classic dualisms falter and fade. This approach reaches metaphysical significance, as close as possible to the texture of reality and just as close to the texture of the mind. Even the most theoretical kind of mathematics can find anchorage in an intuitive kind of geometry and the most virtual algorithm can reveal the indices of the mind’s structure.
By avoiding both the pitfall of a totalising theory and that of indulgence towards technoscience, the exhibition creates a middle way that invites our gazes to cross paths at a time when disciplines are becoming compartmentalised and borders closing up. Yet the approach remains nuanced. Artificial intelligence can turn out to be as much a stone of madness as a Pandora’s box. Not to mention that thinking about AI means finding the link to introspection in order to question what still eludes us, where consciousness resists examination.
Céline Boisserie-Lacroix
Céline Boisserie-Lacroix is a researcher in philosophy and holds a doctorate from the Institut Jean Nicod (EHESS-ENS-CNRS).
Marina Gadonneix (1977) lives and works in Paris. A graduate of the École Nationale Supérieure de la Photographie in Arles, Marina Gadonneix has won numerous awards (HSBC Prize, Dummy Book Award from the Luma Foundation and Rencontres d’Arles, Nièpce Prize, Lewis Baltz Prize). She has been represented by the Christophe Gaillard Gallery since 2017. Her work was featured in a joint exhibition with Lynne Cohen in 2023 at the Centre Pompidou and has been exhibited in numerous institutions in Europe and the United States, including the Fotohof in Salzburg, the Centre Photographique d’Île de France, the Kunsthalle in Tübingen, the Rencontres Photographiques d’Arles and MOMENTA | Biennale de l’image at the Musée de Joliette, Canada. Her press coverage includes Le Monde, Libération, Télérama, Artpress, Le Journal des arts, Etudes photographiques, Canadian Art, and more.
A former student of the École Polytechnique (2017), Victor Rambaud began his career with a major American player in artificial intelligence. After several years in this industry, he turned to cognitive science research to begin a thesis at the ENS on the development of computational models of the mind. It was through a collaboration with artist Lucien Murat that he became interested in the dialogue between art and artificial intelligence, prompting him to reflect on the duality of generated images – on the one hand, their sensitive texture and, on the other, the algorithmic structure on which they are based.
