DAVIDSON’S TRIANGULATION AND ARTIFICIAL INTELLIGENCE: SEMIOTICS OF DIGITAL SUBJECTIVITY

Abstract

The article explores the transformation of Artificial Intelligence (AI) from a mere instrumental tool into an autonomous mediator through the lens of Donald Davidson’s epistemological concept of triangulation. The relevance of the study is driven by the necessity for a philosophical reflection on Large Language Models (LLMs), wich presuppose considering them not merely as algorithmic systems but as potential subjects of communication. This shift requires a re-evaluation of propositional attitude categories, such as beliefs, intentions, and desires. The scientific novelty lies in the juxtaposition of Davidsonian "triangular externalism" (Self–Other–World) with contemporary issues of digital subjectivity. By referencing the "Swampman" thought experiment, the author argues that without historical genesis and a causal connection to a shared world, AI risks remaining a "semantic zombie." In contrast to W. Quine’s approach, Davidson’s strategy asserts that objectivity of knowledge arises only through intersubjective interaction; consequently, understanding truth becomes possible solely through language as a form of social triangulation. Within the scope of the study, a parallel is drawn with the fictional image of the robot Klara from Kazuo Ishiguro's novel as a late model of Davidsonian triangulation, where ethical subjectivity emerges through an attempt to interpret the human context. Particular attention is paid to the logical structure of radical interpretation and the adaptation of Alfred Tarski’s theory of truth. The author demonstrates that the meaning of sentences in AI systems remains "semantically opaque" without engagement in a shared symbolic space. The conclusions emphasize that the transition from mobile to ambient computing requires AI to possess more than just precision; it demands the capacity for the "cognitive utilization of the three-way link." It is established that attributing propositional attitudes to AI is often a result of human anthropomorphism, whereas genuine subjectivity requires a history of interaction with humans. The article proposes regarding triangulation as a fundamental model for developing human-machine systems, where ethical responsibility and meaningful communication become possible only within the "closed triangle" of a shared objective reality.