Quantifying Meaning: A Review of Semantic Information Analysis, Uncertainties, and Possibilistic Restrictions in Hybrid Systems
Abstract
Classical information theory, originating with Shannon’s foundational work [1], treats information as a purely syntactic quantity divorced from meaning. Intelligent systems operating in real-world environments must nonetheless process semantic content carried in natural language — content permeated by vagueness, fuzziness, ambiguity, and multimodal uncertainty. This article presents a structured review of the conceptual and mathematical landscape surrounding semantic information theory, proceeding from Shannon’s technical framework through Weaver’s three-level communication model, Floridi’s strongly semantic theory, Zadeh’s restriction-centred view of natural language, and contemporary rate-distortion-based semantic communication theory, toward a possibilistic approach suited to modern human-machine hybrid systems. The review is organised along four axes: (i) a rigorous taxonomy of uncertainty types along the dimensions of nature, object, and severity; (ii) a hierarchy of Levels Of Abstraction (LOA) linking raw sensor data to high-level semantic content; (iii) a critical, comparative appraisal of classical and contemporary methodologies, including the Bar-Hillel–Carnap paradox, Floridi’s veridical resolution, and recent synonymous-mapping semantic communication theory; and (iv) the Natural Possibility (NP) Framework, a data-driven possibilistic account that converts unimodal frequency data into natural possibility distributions via a modal conversion rule, and defines workable measures — Proximity Value (PV), Elemental Proximity Value (EPV), Mean Proximity Value (MPV), and a semantic measure SM — for precisiating vague human commands in machine reasoning. The article identifies a specific gap in the existing literature — the absence of an operational bridge linking uncertainty taxonomy, the LOA hierarchy, and computable possibilistic measures — and evaluates the extent to which the possibilistic account addresses it, alongside its limitations relative to contemporary coding-theoretic semantic communication frameworks. Open research directions in approximate reasoning, explainable AI, and 6G semantic communication are identified.