A Rule-Based Pipeline for Digitizing and Mapping Karluk Lexical Units in Uzbek Dialects
Year: 2026
Type: conference-paper
Abstract: This paper presents a transparent rule-based pipeline for converting heterogeneous Uzbek dialectological records into normalized, linguistically annotated, and georeferenced data for the areal analysis of Karluk lexical units. The workflow comprises source-preserving transcription, Unicode normalization, deterministic matching, three-criterion linguistic classification, WGS84-based geographic linking, targeted expert validation, and map generation. The pilot dataset contains 57 lexical concepts represented by 260 dialectal variants, corresponding to an average of 4.56 variants per concept. A sample of 20 variants, representing 7.7% of the dataset, was independently audited by a linguist specializing in Uzbek dialectology who had not participated in the initial classification. Sixteen entries required no correction (80%), while four required revision (20%): one dialect-origin reassignment and three restrictions of geographic scope. In an exploratory two-sided exact binomial test, the observed no-correction proportion differed from an equal-probability benchmark of 0.50 (p=0.0118); this result is interpreted as targeted quality-control evidence rather than as full-corpus accuracy or inter-annotator agreement. The computational subset covers 10 of Uzbekistan’s 14 first-level administrative units (71.4%). An illustrative settlement-level map derived from the broader dialectological corpus demonstrates the intended spatial output, while all findings are explicitly presented as exploratory rather than as a complete national dialect atlas or a trained NLP model.
Authors: Sodiqjon Yoqubov, Gulnoza Salimova
Language: English
Subfield: Language and Linguistics
Field: Arts and Humanities
Domain: Social Sciences
Open Access status: hybrid
OpenAlex ID: W7224907603