Send message to Author

Short abstract:

This paper moves across three axes to examine 1. the ontological and epistemological stances re-made at the juncture between robotics and neuroscience;2. the power of, and need for, modelling in sustaining such project; 3. how such alliances are made successful by obscuring critical aspects of life.

Long abstract:

This paper draws from fieldwork in a robotics laboratory aligned with recent neuroscientific developments – the Free Energy Principle. FEP has gained increasing traction, growing from modest hypothesis to elucidate visual processing phenomena, to ‘theory of everything’ aiming to explain life.

I first elucidate FEP's intricacies, highlighting its connections to cybernetics. I suggest that the cybernetic project brought forward by FEP is radically different from the optimistic picture highlighted by Pickering (2010), rather configuring information as a statistically knowable object.

Secondly, I argue the principle’s explanatory power resides in its formalism, generating models which reduce life to causative processes. This formalism can be leveraged by roboticists, who can replicate them algorithmically, enhancing their ontological legitimacy. Despite this pragmatic alliance, I describe an obfuscated clash: the brain-machine entanglement is partial, since engineers look to models as tools to increase efficacy, rather than ontological proofs. Such different interests however often become blurry, as evidenced by an ethnographic vignette of a failed experiment. The objectivistic allure of modelling led roboticists into attributing failure not to a fundamental onto-epistemological fallacy, but rather to limitations of their implementation.

I thus highlight an inability on both sides to recognise modelling practices as producing vital models (Mahfoud et al., 2017). Instead, they push for a seductive, but dangerous, vision of life as it is produced by such modelling practices. By drawing on Simondon (2020) and Kauffman (2019), I conclude by outlining what it is that gets downplayed through such mind-machine entanglements mediated and made possible by modelling.