The Painted Mountain : A Comparative Study of the Cultural Logics in Ink Wash Paintings of Huangshan and Oil Paintings of Yosemite

This thesis develops a corpus-based approach to the comparative study of landscape painting by constructing two curated image corpora: Huangshan paintings (17th century) and Yosemite paintings (19th–early 20th centuries). The selection of these periods is historically significant, as each corpus represents a distinctive moment within its respective painting tradition and cultural discourse of nature. Moving beyond close reading of a small number of artworks, the thesis adopts a corpus-level “distant viewing” strategy. This study examined a sample of 138 artworks, with 69 in each corpus. The analysis results were subjected to cultural interpretation by integrating computer vision feature extraction with manual pattern annotation methods and combining them with iconographic and semiotic theoretical frameworks. Quantitative measures identify systematic differences in visual grammar between the two corpora, while motif statistics summarize the frequency and co-occurrence patterns of annotated categories across datasets. The thesis also evaluates the limits of applying a unified annotation scheme across visual systems with different semiotic structures. Inter-annotator agreement indicates that spatial consensus can coexist with categorical disagreement for ambiguous motifs.