Anderson's Method: A Documented Protocol for Deriving AI-Assisted Artworks from Historical Primary Sources

This article describes Anderson's Method, a research-and-production protocol in which AI-assisted artworks are derived from the close reading of individual historical primary sources — Latin treatises, early modern scores, and the archival documentation of baroque theatrical engineering. The method is positioned as a generative extension of media archaeology: where the archaeological tradition excavates alternative genealogies of technical media analytically, this protocol converts excavation into production while preserving a verifiable chain from source page to finished artifact. The article gives operational definitions of the method's core terms (filter, structure, trail), specifies four extractable structure types, distinguishes two modes of AI participation with a mandatory verification step, presents a complete traced example, and addresses the apparent contradiction between the method's declared subjectivity and its claim to verifiability. Subjectivity, it is argued, is not a defect to be engineered away but a documented parameter of the system.