Skip to content
DSML.KZ
Publications

Talk · International Conference on Computer Science and Engineering · September 1, 2026

An Audited Bilingual Corpus and Annotation Protocol for the Mother Archetype in Russian and Uzbek Women’s Prose

Дилдора Муродова, Gulnoza Salimova

Gulnoza Salimova
Open via DOI

Abstract

This paper presents an audited bilingual corpus and an annotation protocol for studying the mother archetype in selected prose by Masha Traub and Zulfiya Qurolboy qizi. The corpus comprises 39,401 word tokens and 4,871 sentences drawn from seven works, and every work is documented individually by edition, source, coverage, script, and analysed length. A corpus audit identifies three properties that materially constrain any inference drawn from this material: the Russian subcorpus is 1.61 times larger by tokens; source completeness is asymmetric, since three Russian books are represented by public excerpts while all three Uzbek stories are complete; and two works alone contribute roughly half of each language. We define a five-category multi-label annotation scheme for maternal sub-roles, and specify an analysis design with the safeguards the audit shows to be necessary: normalised proportions with stated denominators, work-level grouping, leave-one-work-out sensitivity analysis, and set-valued agreement measurement. A pilot annotation of a balanced 80- sentence sample (40 Russian, 40 Uzbek) by two independent coders yielded a raw agreement of 75.0 %, Cohen’s κ = 0.662 and Krippendorff’s α = 0.668, indicating substantial reliability of the scheme. While the paper intentionally defers full- corpus sub- role estimates, this pilot demonstrates that the annotation framework is operational and can be applied to the complete audited corpus once the annotation is finalised.

Digital Humanities and ScholarshipNatural Language Processing TechniquesLinguistic research and analysismother archetypematernal rolewomen's prose

More by Gulnoza Salimova