Aligning the Odes: A Character-Level Parallel Dataset for the Anhui University and Fùyáng Bamboo-Slip Manuscripts

This dataset aligns graph-level transcriptions of three witnesses to the Book of Odes: the received Máoshī 毛詩, the Warring States so-called ‘Anhui University’ (Āndà 安大) bamboo manuscript, and the Western Han Fùyáng 阜陽 bamboo manuscript. The release represents each witness in Unicode JSON, uses Ideographic Description Sequences for graphs without suitable encoded characters, and records explicit one-to-one, one-to-many, and many-to-many correspondences. Separate mappings connect transcription positions to physical locations in the source editions and those locations to images of the manuscript characters. JSON Schemas, integrity tests, a static read-only viewer, and reviewed crop recipes support validation and reproducibility. The corpus makes manuscript variation available both as structured data and through a visual interface in which readers can inspect aligned passages, graph images, and their positions on reconstructed manuscript strips. Its format can accommodate later manuscript publications and competing transcriptions of the same physical witness. It turns manuscript and transcription variation into an inspectable, reusable research infrastructure for early Chinese textual transmission, palaeography, historical phonology, and digital philology.