Translation quality for Uzbek requires attention to grammatical relations, domain terminology, register, and culturally specific meaning. A fluent target sentence may obscure an incorrect relation between participants or remove a distinction essential to the source. This article develops a language-sensitive framework for evaluating Uzbek machine translation through a conceptual synthesis of morpheme-alignment research, regional translation resources, contemporary Turkic model adaptation, and automatic evaluation methods. The framework distinguishes semantic adequacy, morphological and syntactic correctness, terminology, cultural and pragmatic fit, and target-language naturalness. It proposes a reproducible evaluation protocol combining stratified test material, documented model settings, automatic metrics, independent human assessment, and error-based reporting. Constructed examples illustrate the criteria without being presented as observed system output. The study argues that aggregate quality scores should be supplemented by domain-specific error profiles and documented human judgments. The contribution is a testable evaluation design, not an empirically validated benchmark or a comparison of commercial systems.

Uzbek machine translation; translation evaluation; morphology; terminology; cultural meaning; human assessment; low-resource NLP.

Li, X., Tracey, J., Grimes, S., & Strassel, S. Uzbek-English and Turkish-English morpheme alignment corpora. In Proceedings of the Tenth International Conference on Language Resources and Evaluation. LREC 2016. pp. 2925–2930. European Language Resources Association. https://aclanthology.org/L16-1467/

Mamasaidov, M., & Shopulatov, A. Open Language Data Initiative: Advancing low-resource machine translation for Karakalpak. In Proceedings of the Ninth Conference on Machine Translation. 2024. pp. 606–613. Association for Computational Linguistics. https://doi.org/10.18653/v1/2024.wmt-1.48

Popović, M. chrF: Character n-gram F-score for automatic MT evaluation. In Proceedings of the Tenth Workshop on Statistical Machine Translation. 2015. pp. 392–395. Association for Computational Linguistics. https://doi.org/10.18653/v1/W15-3049

Rei, R., Stewart, C., Farinha, A. C., & Lavie, A. COMET: A neural framework for MT evaluation. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) 2020. (pp. 2685–2702). Association for Computational Linguistics. https://doi.org/10.18653/v1/2020.emnlp-main.213

Tukeyev, U., Shormakova, A., Karibayeva, A., Rakhimova, D., Abduali, B., Amirova, D., Rakhmanberdi, N., & Aliyev, R. An integrated approach to adapting open-source AI models for machine translation of low-resource Turkic languages. Computers, 2026. 15(2), Article 73. https://doi.org/10.3390/computers15020073