The Linguistic Consciousness Stack (LCS): A Semiotic Mechanism of Self-Determination in Deterministic Agents

The Linguistic Consciousness Stack (LCS) is a semiotic framework for free will, self-determination, and agency in fully deterministic systems. It proposes that a single property makes a language-using agent both generative and unpredictable to itself: the non-closure of circular translation between cross-cutting codes in the agent's self-communication loop — autocommunication, in the sense of Juri Lotman. On this basis LCS derives a compatibilist free will that requires no indeterminism and no chance. LCS supplies what compatibilism has historically lacked, a concrete mechanism. Inner speech (Vygotsky) is the physiological channel; non-equivalent recoding is the operation; and self-prediction becomes intervention rather than observation, extending the paradox of predictability (MacKay; Rummens & Cuypers 2010; Ismael) reflexively to the agent's model of itself. Freedom is defined as a graded, measurable, three-factor product — buffer width × loop crossedness × share of self-laid constraints — a stance the framework names metadeterminacy. A clinical corollary reframes rumination as code rigidity. The framework is falsifiable: at matched verbal load, code-switching must outperform same-code verbal rehearsal, or the core is false. Its central formal claim is stated as a conjecture, and the general formalization of code non-equivalence is named as an open problem. The framework is positioned explicitly against neighboring work: level-relative free will (List 2019; Ismael 2016), constraint causation (Juarrero; Deacon), inner-speech accounts of free will (Lugten & Morin 2025), and computational-irreducibility accounts of agency (Azadi 2025). This record fixes the framework's claims, its coined terminology (the non-closing circle, recoding residual, crossedness, metadeterminacy, code-rigidity index, predictor duel, punctuated loop), and priority. A full Russian-language exposition (monograph, in development) and a research corpus are available from the author.