Interpretable Psycholinguistic Profiling of Uzbek Student Essays via a Minimal Lexicon
Abstract
The article presents an interpretable algorithm for profiling student essays in the Uzbek language based on a minimalistic psycholinguistic lexicon. The lexicon is organized into seven psychologically motivated blocks (self/social reference; cognitive processes; deontic/epistemic modality; affectivity; social/value markers; temporality/goal setting; discourse connectives). An algorithm implemented on the basis of a rule-oriented approach is proposed. Testing was conducted on a corpus of 50 essays from various fields (economics, jurisprudence, etc.). The obtained data confirm the methodological feasibility of simultaneously considering both reports: the first reflects the prevalence of markers, the second - their intensity. Limitations (genre dependence, sensitivity to irony/metaphor) are discussed, and directions for development are outlined: expanding the lexicon, deepening the rules for processing negations, and embedding lightweight syntactic templates for cause-and-effect structures. In addition, for a more detailed understanding of the complexities and, in general, problems of processing Uzbek texts, the authors described in detail the features of the Uzbek language, its nature and morphology in general.