Some semiotic dimensions of generative AI: Structure, generativity, and relational epistemology

The common use of ‘generation’ may suggest an epistemic link between Chomskyan generativity, the Greimasian parcours génèratif, and modern systems of Generative AI. This article, however, reconstructs the relationships between contrasting epistemologies – specifically, the structural semiotic approach vs. Chomskyan generative syntax – to distinguish and semiotically analyse AI-driven textual generation. While the Chomsky-Greimas relationship has been extensively studied and debated, its relevance to the epistemological presuppositions of General Purpose Chatbots (e.g., ChatGPT, Gemini) remains underexplored. Generative AI's operation thus provides an empirically observable basis for semiotic analysis and reflection on language, relaunching the semiotic project to address challenges arising from the widespread use of these linguistic technologies. This study also reaffirms the profound semiotic-anthropological relevance of Greimasian generativism concerning sense as it is structured in the lived world, as well as the importance of not separating semantics from pragmatics, as indicated by Umberto Eco’s interpretative framework of semiotics.