Supplementary Data for Two Millennia of Moral Language: Semantic Stability and Change in Chinese Historical Texts

This record contains the compact derived-data and reproducibility package for the manuscript Two Millennia of Moral Language: Semantic Stability and Change in Chinese Historical Texts. The study examines semantic stability and change in moral-related Chinese lexical forms across transmitted written Chinese from 221 BCE to 1912 CE. The archive includes the fixed lexeme and matched-control inventories; temporal and PPMI comparison tables; coverage and date-curation records; 87 author-coded readings; 16 purposive historical passage comparisons; source data for figures and tables; analysis specifications; checksums; and code for verifying the package and its key result inventories. The corpus census covers 2,200,039,665 Han characters before cross-source deduplication, while the released archive contains derived evidence rather than copies of the source corpora. Full corpora, cached occurrence windows, and contextual embedding arrays are not redistributed because of scale and the original providers' terms. The archive preserves eligible, ineligible, and negative-result comparison rows so that the reported evidence can be inspected rather than reconstructed from selected outputs alone. Licenses: data, documentation and metadata under CC BY 4.0; software in code/ under the MIT License (see code/LICENSE).