Lexical Diversity, Structural Transformation, and Collocational Features in a Six-Language Parallel Subcorpus

This study analyzes lexical diversity, structural transformation, and collocational features based on a six-language parallel subcorpus consisting of Kazakh, English, Uzbek, Turkish, Uyghur, and Azerbaijani texts. The main objective of the research is to identify lexical and structural changes occurring during translation across the six languages using a corpus-based approach and to demonstrate the typological similarities among Turkic languages from a comparative perspective. A multilingual parallel corpus organized according to the principle of paragraph-level alignment was used as the research material. The corpus data were processed in the R programming environment, where lexical diversity analysis, structural transformation analysis, and bigram-based collocation analysis were conducted. To determine lexical diversity, total words, unique words, and type-token ratio (TTR) values were calculated. Structural transformations were identified through three translation models: expansion, compression, and stable structures. In addition, collocation structures across the languages were compared through bigram analysis. The results demonstrated relatively high lexical diversity indicators in the Turkic languages. The highest TTR value was identified in Turkish (0.381), while English showed the lowest result (0.156). Structural transformation analysis revealed that English translations tend toward expansion strategies (69.5%), whereas Turkish translations demonstrated a high proportion of compression structures (46.0%). Collocation analysis further showed that character-centered and dialogue-based structures are preserved at a high level across the Turkic languages. Overall, the findings confirm that multilingual parallel corpora constitute an important resource for translation studies, corpus linguistics, and comparative linguistics.