Key Semantic-Domain Profiles in Corpus Samples Retrieved Using English black and Chinese 黑: An Exploratory Corpus-Based Analysis
Abstract
This exploratory corpus-based study examines the key semantic-domain profiles that emerge in samples retrieved using English black and Chinese 黑, the USAS code-level overlap between the two complete lists, the thematic contrasts among the highest-ranked non-Z items, and the interpretive effect of an alternative English reference comparison. The English data comprise 10,000 sampled text records each for black and white from the British National Corpus, while the Chinese data comprise 6,000 sampled text records each for 黑-related and 白-related forms from Chinese Web 2011. Wmatrix results are interpreted within each language and then synthesized at the list level. Relative to their respective reference samples, the English black sample foregrounds people, groups, public discourse, the universe, and the economy, whereas the Chinese 黑-related sample foregrounds colour, crime, materials, and technology. The complete lists share only three USAS code-level overlaps—S5+, E3-, and A5.1-. With the BNC 1994 Sampler Written as the alternative English reference corpus, the domain Colour and colour patterns ranks first, suggesting that comparison baselines and reference-corpus composition may affect domain visibility. By combining separate within-language comparisons, code-level overlap, a limited synthesis of the highest-ranked non-Z items, and the alternative English comparison, the study offers a transparent, reference-sensitive account of these profiles. They are interpreted as relational characteristics of retrieved samples, not as fixed lexical senses or cultural facts.