This white paper/preprint proposes a constraint-structured model of linguistic meaning as a formal research architecture for computational linguistics and AI. It translates a psycholinguistic model of linguistic consciousness into computationally explicit terms by representing interpretation as dynamic interaction among four coupled semantic functions: denotation, designation, connotation, and consignation. These functions are modeled within a bounded semantic state space shaped by linguistic-form and referential systems, and are further open to contextual, cultural, institutional, and sensory inputs. The paper argues that explicit internal semantic organization may improve interpretability in distributional and hybrid AI systems, especially in semantically layered tasks such as cross-linguistic translation, historically sedimented vocabulary, legal and institutional discourse, heritage-language materials, and culturally differentiated lexical fields. It further suggests that, if aligned with data on individual and group language practices and sensory input regimes, the same principle may contribute to individually differentiated deep learning. The present version is a preprint/white paper research document. It includes formalization, visual modeling, and an operational research agenda intended for computational linguists, AI researchers, and interdisciplinary teams working on semantic architecture, language modeling, and culturally responsive AI.
Gayane Hovhannisyan (2026) studied this question.