Provenance, Collaboration, Bibliography, and Value: A Network Analysis of the National Gallery of Art

Abstract

This thesis contributes to digital art history and the emerging field of cultural heritage network analysis in cultural heritage by applying network quantitative measurement, network visualization, and macroscopic interpretation to the permanent collection of the National Gallery of Art in Washington, D.C. The corpus consists of approximately 140,000 artworks for which no known comprehensive network study has previously been undertaken. Grounded in a complex network framework, this analysis treats collections as emergent, socially produced systems of relations, drawing on distant reading and graphic semiology to render institutional structure legible at scale. The research interrogates how provenance, collaborative interactions, and bibliographic citation shape the connective tissue of a national collection. The question of collection value runs parallel to this analysis, reframing value as one among several partially emergent relational properties that network structure can possibly illuminate. Three networks are constructed to distinguish the institutional structure: a constituent-constituent chain network, a collaborator co-occurrence network, and a constituent-bibliography network. Metrics of network centrality were applied to several sub-networks in order to explore the data topology, including betweenness, harmonic closeness, and eigenvector centralities. The analysis finds that the NGA's collection is topologically a composition of donor sub-collections rather than a single unified holding, the Systematic Catalogue volumes dominate bibliographic attribution to create institutional anchor points for American and British painting, and Baroque paintings producing the longest explicit provenance chains with up to twenty-eight recorded owners. Each network is additionally examined through the lens of collection valuation for what it can and can not show regarding dimensions of value.