Semantic category theory cognition and conceptual reference

Abstract

Purpose

Identifying the fundamental characteristics of meaning and deriving an automated meaning‐analysis procedure for machine intelligence.

Design/methodology/approach

Semantic category theory (SCT) is an original testable scientific theory, based on readily available data: not assumptions or axioms. SCT can therefore be refuted by irreconcilable data: not opinion.

Findings

Human language involves four totally independent semantic categories (SC), each of which has its own distinctive form of “Truth”. Any sentence that assigns the characteristics of one SC to another SC involves what is termed here “Semantic Intertwine”. Semantic intertwine often lies at the core of semantic ambiguity, sophistry and paradox: problems that have plagued human reason since antiquity.

Research limitations/implications

SCT is applicable to any endeavour involving human language. Research applications are therefore somewhat extensive. For example, identifying metaphors posing as science, or natural language processing/translation, or solving disparate paradox types, as illustrated by worked examples from: The Liar Group, Sorites Inductive, Russell's Set Theoretic and Zeno's Paradoxes.

Practical implications

To interact successfully with human language, behaviour, and belief systems, as well as their own environment, intelligent machines will need to resolve the semantic component/intertwines of any sentence. Semantic category analysis (SCA), derived from SCT, and also described here, can be used to analyse any sentence or argument, however complex.

Originality/value

Both SCT and SCA are original. Whilst “category error” is an intuitive notion, the observably precise nature, number and modes of interaction of such categories have never previously been presented. With SCT/SCA the rigorous analysis of any argument, whether foisted, valid, or obfuscating, is now possible: by man or machine.

Citation

St Quinton, J.G. (2005), "Semantic category theory cognition and conceptual reference", Kybernetes, Vol. 34 No. 9/10, pp. 1321-1348. https://doi.org/10.1108/03684920510614704