Functional-Semantic, Cognitive, and National-Cultural Representation of the Lexicon Denoting Human Psychological States and Its Computational-Linguistic Modeling
Year: 2026
Type: conference-paper
Abstract: This study examines how psychological states are expressed in Uzbek and how these expressions can be represented in computational models. Particular attention is given to lexical and phraseological units whose meanings depend not only on emotion itself but also on cultural and social context. Words such as andisha and uyat, for example, are difficult to describe through simple positive-negative polarity because they also reflect socially regulated behaviour and cultural expectations. Drawing on functional-semantic, cognitive, and linguocultural analysis, the study develops a pilot Uzbek emotion lexicon in which lexical units are described according to emotion category, valence, arousal, and a proposed Cultural Weight parameter. The paper also outlines a hybrid emotion-detection approach that combines lexicon-based information with contextual transformer representations. Muxlisa AI is considered as a possible application environment for this framework. Preliminary correlation results suggest that Cultural Weight may capture information that is not fully represented by valence and arousal; however, these findings relate to the pilot annotation framework rather than to the performance of a fully trained emotion-classification model.
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7224771693