The Coloniality of Metadata: A Critical Data Analysis of the Archive of Early American Images at the John Carter Brown Library

How do contemporary metadata practices replicate the coloniality of power embedded in European colonial documents describing the Spanish Americas? This report draws from critical theory within the field of Latin American studies to explore the extent to which standardized description and categorization practices can perpetuate a Eurocentric colonial gaze on the Spanish Americas. In order to ground this theoretical engagement within the fields of information science and critical data studies, a dataset is compiled and computationally analyzed, which contains metadata records corresponding to images derived from books, manuscripts, and broadsides printed between 1492 and 1825 about the Spanish Americas found in the Archive of Early American Images at the John Carter Brown Library. This report then applies unsupervised machine learning techniques such as Latent Dirichlet Allocation (LDA) topic modeling, and k-means clustering, and counts word frequencies to identify broad metadata trends across the collection. Ultimately, this report combines methods from critical data studies and Latin American cultural studies to begin to understand how controlled vocabularies and descriptive practices can perpetuate colonial structures of power.