Machine, organism and language: a comparative epistemology of AI models

Abstract Inspired by Ernst Cassirer and Michel Foucault, this essay proposes a comparative epistemology of three paradigms central to the making of modern science, the humanities, and more recently AI: machine, organism, and language. These paradigms have influenced one another and recombined into complex analogies. Whereas the philosophy of science has often emphasised the organism-machine analogy from early modern mechanicism to cybernetics, this essay extends the inquiry to the language-machine analogy of late modernity, which runs from the telegraph and the Turing machine to information theory and Large Language Models. The rise of AI is thus framed as the confluence of these three paradigms, read not from an internalist perspective but from an externalist one, as mirrors of the social order. Against the dominant view of AI as a purely mathematical achievement or an imitation of biological intelligence, the essay argues that what AI systems automate are the relational structures sedimented in human cooperation, the division of labour, and culture at large—making AI, in effect, a model of the social manifold.