You do not currently have permission to view the full text for this document. Please sign in to an account with an active subscription or contact support.
This text investigates the effects of uncanniness provoked by the functioning of large language models (LLMs), articulating them with Freud’s concept of the uncanny (das Unheimliche) and Lacanian elaborations on the structure of language. The starting point is the observation that, although Lacan extensively mobilized the concept of repetition compulsion in the context of machines, he curiously left aside the notion of the uncanny, despite its conceptual and temporal proximity to repetition. We argue that LLMs actualize a new form of uncanniness, no longer grounded in the humanoid appearance of robots, but in the emergence of meaningful linguistic combinations produced by a language without body and without world. The formal and statistical structure of LLMs highlights the machinic nature of language, revealing something unsettling about the very functioning of natural language and meaning: their production can occur autonomously, without reference to subjective intention or lived experience. By exploring this hypothesis, the text proposes that the experience of uncanniness elicited by LLMs not only signals a breakdown in the distinction between natural and artificial language, but also exposes the constitutive artificiality of human language itself.